Faculty Recruiting Support CICS

Scaling Down the Energy Cost of Connecting Everyday Objects to the Internet

15 Oct
Tuesday, 10/15/2019 9:30am to 11:00am
Ph.D. Dissertation Proposal Defense


Internet of Things (IoT) has seen a dramatic growth over the past few years. This popularity means that small objects that connect via wireless to the internet are appearing in different applications and use cases. However, extending IoT to different applications needs to address serious fundamental gaps in terms of energy between these devices and commodity wireless infrastructure. At the highest levels, commodity radios are not energy-optimized and thus need to be replaced with passive radios which better fit the limited energy budgets of IoT devices. On the other hand, existing passive radios are not compatible with commodity infrastructure as well as have a very poor robustness. In this thesis, we focus on different scenarios that these gaps exist and try to find new paradigms for resolving them.

We first introduce the idea of radio polymorphism, which means combining passive radios such as backscatter (transmitter) and envelope detector (receiver) with some active elements to make them more reliable. We realize that passive radios cannot maintain their connection in scenarios that involve continuous streaming of data from low power devices to a bigger device or a cloud server. In this scenarios, active elements can reinforce the performance whenever it is needed to do so. Polymorphic radio design is beyond a simple active-passive combination and various techniques have been adopted so that the radio can dynamically and intelligently adapt to channel conditions, the type of application, and so on.

Our next study is about the viability of having battery-free RFID tags that can directly connect to commodity devices such as WiFi routers, cell-phones, etc., as opposed to standard RFID tags which need to deploy dedicated devices, called RFID readers. We show the viability of this goal by designing and implementing a backscatter tag, called xSHIFT, that can collect the required energy as well as the activation signal from WiFi devices and transmit a backscatter signal that can be decoded by commodity Bluetooth receivers. This allows us to attach battery-free tags to different objects and products in homes, offices, stores, etc. and scan them without any additional infrastructure.

Finally, our ongoing project focuses on designing a more robust receiver for ultra low power devices. Traditional ultra low power receivers which are based on a simple envelope detector, while being idea in terms of power consumption, are very  poor in converting the received RF signal to baseband information. So, their operation is limited to a few meters range at best. We propose a new technique that boosts the receive range by several times while keeping the receiver micro-power. This idea is based on helping the envelope detector by feeding an external helper carrier signal externally. Envelope detector in this case mixes the received signal with this helper carrier which results in a much better conversion gain.

Advisor: Deepak Ganesan